Projected End Date: September 2014
Objectives
Develop a flexible and scalable computational framework that can integrate multiple physics models at various scales (battery pack, cell, electrodes, etc.), and provide a predictive modeling tool under the auspices of the CAEBAT program.
Coordinate with partners across the program on requirements and design of the framework so as to preserve the investment in existing models.
Ultimately, the detailed simulation capability will model coupled physical phenomena (charge and thermal transport; electrochemical reactions; mechanical stresses) across the porous 3D structure of the electrodes (cathodes and anodes) and the solid or liquid electrolyte system while including nanoscale effects through closures based on resolved quantities. The simulation tool will be validated both at
the full-cell level and at the battery-pack level, providing an unprecedented capability to design next-generation batteries with the desired performance and safety needs for transportation.
Technical Barriers
Given the complex requirements for development of electrical energy storage devices for future
transportation needs, a predictive simulation capability which can guide rapid design by considering
performance and safety implications of different chemistry and materials choices is required. This capability must leverage existing investments and integrate multiple physics models across scales in order to (1) provide feedback to experiments by exploring the design space effectively, (2) optimize material
components and geometry, and (3) address safety and durability in an integrated fashion. Such models do not currently exist.
Technical Targets
Develop the computational framework that will integrate existing models and new models developed by different CAEBAT subcontractor teams that span across the battery pack, modules, cells, etc. to provide an integrated design tool to battery manufacturers to optimize performance and safety in an accelerated fashion.
Accomplishments
Released Beta V1a of the CAEBAT-OAS framework together with VIBE (Virtual Integrated Battery Environment), Battery ML (BatML) Schema specifications, battery state, and few examples.
Completed porting of OAS (Open Architecture Software) to Windows.
Revisions to the BatML standard and translators to/from: ANSYS, EC-Power, and AMPERES.
Integrated workflow environment through NiCE: job launch, postprocessing of the results, XML files editing.
Introduction
Computational tools for the analysis of performance and safety of battery systems are not currently
predictive, in that they rely heavily on fitted parameters. While there is ongoing experimental research at various length scales around the world, computational models are primarily developed for the lower-length scales (atomistic and mesoscopic), which do not scale to the system-level. Existing models at the macroscopic or system-level are based on electrical circuit models or simple 1D models. The 1D models are limited in their ability to capture spatial variations in temperature, potential in the electrical circuits of the battery cells and
IV.C.2 Computer Aided Engineering of Batteries Turner – ORNL
packs. Currently there is no design tool for batteries that can leverage the significant investments in modeling efforts across DOE and academia. An open and flexible computational framework that can incorporate the diverse existing capabilities and new capabilities coming through CAEBAT partners, can provide a foundation for a predictive tool for the rapid design and prototyping of batteries.
Approach
We are developing a flexible, robust, and
computationally scalable open-architecture framework that integrates multi-physics and multi-scale battery models. The physics phenomena of interest include charge and thermal transport, electrochemical reactions, and mechanical stresses. They operate and interact across the porous 3D structure of the electrodes (cathodes and anodes), the solid or liquid electrolyte
system and the other battery components. The underlying lower-length processes are accounted for through closure equations and sub-models that are based on resolved quantities. The schematic of this framework is given in Figure IV - 63.
The end result will be a verified, computationally scalable, portable, and flexible (extensible and easily- modified) framework that can integrate models from the other CAEBAT tasks and industrial partners. The framework will be used to validate models and modeling approaches against experiments and to support rapid prototyping of advanced battery concepts. Figure IV - 64 provides the roadmap for initial loosely- coupled model integration framework with a fully- implicit coupled capability in the later years.
Figure IV - 63: Schematic of the OAS modeling framework and interactions with other tasks within the CAEBAT program and external activities.
Turner – ORNL IV.C.2 Computer Aided Engineering of Batteries
Figure IV - 64: Coupling scenarios in battery modeling. We started with one-way and two-way loose coupling. In later years, as needed, we will move towards two-way tight coupling with Picard and Full-implicit methodologies
Results
Virtual Integrated Battery Environment (VIBE).
Integration of several components (pseudo-2D DualFoil, NTG, AMPERES, NREL’s MSMD) has been completed. Initial linking to the ANL cost model has been done. In this current scenario the electrochemical component in VIBE supplies the area specific impedance (ASI) to the cost model to be further used in battery parameters calculations. The results of the modeling of a pouch cell (Farasis Energy, Inc.) were validated by experimental measurements of the cell surface temperature during discharge (Figure IV - 65). Excellent correlation can be observed that provides confidence in the modeling approach and integration of components in OAS.
Flexibility of the OAS was tested by substitution of one of the components in VIBE (DualFoil) with another (NTG) for electrochemical modeling. It was determined that with finer discretization of the electrodes in DualFoil the results from the two models are nearly identical. This provides users with a choice of the model
most suitable for particular simulation scenario. DualFoil can be used when the details of concentration across the cell sandwich are needed, while NTG can be used when the thermal analysis is the primary goal in addition to significant savings of compute time.
Module level coupling allowed performing simulations of modules consisting of four pouch cells connected either in parallel or in series. Simulations of uneven cooling conditions on the module surface show that the potential difference in the cells on two sides can be as high as 2.5 mV. The battery state was expanded in order to include depth of discharge as an additional variable passed between the components. Figure IV - 66 shows the temperature distribution in a module with four cells in parallel. Initial integration of the mechanical component in VIBE has been performed. Coupling with mechanical modeling including elastic and elastic-plastic response of the material allows simulating scenarios involving battery abuse (Figure IV - 67) and provides guidance for battery safety testing. One-way Coupling Two-way Loose Coupling Two-way TightCoupling s e s e c c , , T t1 t2 s e s e c c , , T s e s e c c , , T t1 t2 s e s e c c , , T s e s e c c , , T t1 t2 s e s e c c , , T Picard self-consistent iterations to some convergence criteria
s e s e c c , , T t1 t2 s e s e c c , , T Fully Implicit Consistency at each iteration across the physics in terms of full non-linear residual
IV.C.2 Computer Aided Engineering of Batteries Turner – ORNL
Figure IV - 65: Validation of 4.3 Ah pouch cell modeling (solid lines) with experimental temperature measurements (markers)
Figure IV - 66: Temperature distribution in a module with assymetric cooling
Figure IV - 67: Mechanical abuse of cylindrical cell
(electrochemical-electrical-thermal-mechanical components) OAS Capabilities of DAKOTA optimization toolkit were explored by running a numerical study of the effect of tab placement on cell temperature. 2000 configurations were run within the simulation with geometry parametrization and automated mesh generation. The lowest temperature was determined in the cell configuration where the tabs were placed on the opposite edges; the effect was more pronounced with an increase of the width of the cooled tabs. Example of
Turner – ORNL IV.C.2 Computer Aided Engineering of Batteries
mesh generation and the temperature distribution corresponding to one of the arrangements is shown in Figure IV - 68.
Figure IV - 68: Tab placement study using DAKOTA
Graphical User Interface and Integrated
Workflow Environment. The development of a tool for
simulation launch and post-processing of the results was based on NiCE project for workflow and data
management. In 2013 we have deployed:
Input editing for OAS setup files. Editing for BatML files.
Local and remote OAS job launch.
Multi-file upload and download of OAS VIBE data
3D static visualization of output.
A screen shot of CAEBAT-NiCE environment is shown in Figure IV - 69. The tool provides easy model setup with drop-down menus for model (component) selection, simulation control parameters and input of the material properties.
Bat ML. The Battery Markup Language (BatML)
supports the CAEBAT OAS and enables standardized generation of simulation input files. As an essential part of the development, translation back and forth to various other native formats should be enabled. In 2013, we completed translators to/from EC Power, ANSYS, and AMPERES. The XML validating tool against BatML schema has been completed and can be used to validate the user supplied XML files. As an example, Figure IV - 70 shows the input file for thermal component in battery simulation translated to BatML format.
IV.C.2 Computer Aided Engineering of Batteries Turner – ORNL
Figure IV - 70: Input file for thermal component (AMPERES) translated to BatML
Conclusions and Future Directions
CAEBAT OAS framework core is stable and has been ported to Windows. Components for
electrochemical, electrical, and thermal modeling have been successfully integrated and initial coupling to a mechanics model has been done. The framework possesses the ability for exchange of the components and integration of DAKOTA optimization toolkit provides a unique set of instruments to perform parametric sweeps and optimization study. Job launch and results post-processing through NiCE gives users an organized and easy to use workflow environment for battery simulations.
In the following year, we will:
Complete integration of the mechanical component in VIBE.
Extend battery state definition to include battery pack simulations.
Implement additional BatML translators as necessary.
BatML revisions based on community feedback.
Release another version of the standard and associated tools.
Implement two-way coupling in OAS. Finalize post-processing and real-time
manipulation in NiCE.
Develop a refined and user-friendly BatML editing in NiCE.
FY 2013 Publications/Presentations
1. Allu, S., Kalnaus, S., Elwasif, W., Simunovic, S., Turner, J.A., and Pannala, S., 2014, “A New Open Computational Framework for Highly-Resolved Coupled Three-Dimensional Multiphysics Simulations of Li-ion Cells,” Journal of Power
Sources, 246, pp. 876-886.
2. Pannala, S., Allu, S., Elwasif, W., Kalnaus, S., Simunovic, S., and Turner, J.A., “Understanding the Effect of Temperature Gradients in Modules on Cell Balance Using Coupled Multi-Physics Modeling Approach With OAS Framework”, ECS
224th Meeting, San Francisco, CA, Oct. 27 – Nov.
1, 2013.
3. Kalnaus, S., Allu, S., Pannala, S., Elwasif, W., Simunovic, S., and Turner, J.A., “Three- dimensional Thermal, Electrical, and
Electrochemical Modeling of Li-ion Batteries,”
ECS 224th Meeting, San Francisco, CA, Oct. 27 –
Nov. 1, 2013.
4. Allu, S., Elwasif, W., Pannala, S., Kalnaus, S., Simunovic, S., and Turner, J.A., “Optimization of tab Placement in Li-ion Battery Using Multi- Physics Simulations,” ECS 224th Meeting, San
Francisco, CA, Oct. 27 – Nov. 1, 2013.
5. Simunovic, S., Allu, S., Pannala, S., Kalnaus, S., Elwasif, W., and Turner, J.A., Coupled Multi- Physics Model for Li-ion Battery Cells During Impact,” EUROMAT 2013 – European Congress and Exhibition on Advanced Materials and Processes, Seville, Spain, Sept. 08 – 13, 2013.